Home on Sydney Data Intelligence Lab

Lu Qin

Mon, 01 Jan 0001 00:00:00 +0000

Wenjie Zhang

Mon, 01 Jan 0001 00:00:00 +0000

Hanchen Wang

Mon, 01 Jan 0001 00:00:00 +0000

Leads the Sydney Data Intelligence Lab. Works on the data management underneath intelligent systems: how they remember, how their parts coordinate, and whether the data underneath holds up.

Yuxin Jin

Mon, 01 Jan 0001 00:00:00 +0000

Works on memory management for self-evolving agents, interfacing with the Memory Engine.

Shulun Chen

Mon, 01 Jan 0001 00:00:00 +0000

Works on latent communication and multi-agent orchestration, interfacing with the Coordination Engine.

Kuiye Ding

Mon, 01 Jan 0001 00:00:00 +0000

Working towards a systems-oriented memory direction.

Siyuan Zhang

Mon, 01 Jan 0001 00:00:00 +0000

Works on retrieval-augmented generation and data prefetching for multi-agent systems, interfacing with the Memory Engine.

Shenghao Xu

Mon, 01 Jan 0001 00:00:00 +0000

Works on agent runtime.

Wei Cao

Mon, 01 Jan 0001 00:00:00 +0000

Algebraic Subgraph Counting

Sun, 13 Jun 2027 00:00:00 +0000

Subgraph isomorphism counting is a fundamental problem in graph analytics, which aims to find the number of subgraph isomorphisms of a query graph in a data graph. The candidate tree-based framework provides a promising foundation for subgraph counting tasks, offering a unified counting paradigm that can be extended beyond tree patterns. However, supporting subgraph isomorphism within this framework remains challenging, as it requires handling both the non-tree edge constraint and the injective mapping constraint. Although existing solutions employ sampling or learning techniques to address these constraints within this framework, they still either suffer from inherent sampling failures or rely heavily on supervision. In this paper, we propose ASC, an algebraic subgraph isomorphism counting approach built on the candidate tree-based counting framework. In our method, the non-tree edge constraint is directly incorporated into the candidate tree-based counting process through a matrix-based computation method, enabling subgraph homomorphism counting with high accuracy in polynomial time. Based on the resulting homomorphism count, we further apply a local sampling method to address the injective mapping constraint, thereby obtaining the final subgraph isomorphism count. Extensive experiments show that ASC can achieve substantially better and more stable performance over the baselines across various datasets, while scaling to billion-edge graphs. Most impressively, as a non-learning method, ASC can even obtain up to over an order of magnitude higher average accuracy than the state-of-the-art learning-based method FlowSC with similar efficiency.

New version of the website is online

Sat, 29 Aug 2026 00:00:00 +0000

A new version of the Sydney Data Intelligence Lab website is online, with our research programme, projects, publications and people in one place.

Elected to the UTS FEIT Faculty Board

Thu, 06 Aug 2026 00:00:00 +0000

Hanchen has been elected as an academic staff member, representing the School of Computer Science, to the Faculty Board in Engineering and Information Technology (FEIT) at UTS. The term commences on 1 January 2027.

FADTI: Fourier and Attention Driven Diffusion for Multivariate Time Series Imputation

Thu, 06 Aug 2026 00:00:00 +0000

UniCom: A Unified Framework for Community Search and Detection via Transfer Learning

Thu, 06 Aug 2026 00:00:00 +0000

Two papers accepted by IEEE ICDM 2026

Wed, 05 Aug 2026 00:00:00 +0000

Two papers accepted by IEEE ICDM 2026 (CORE A*): “UniCom: A Unified Framework for Community Search and Detection via Transfer Learning” (congratulations to Yifan) and “FADTI: Fourier and Attention Driven Diffusion for Multivariate Time Series Imputation” (congratulations to Runze).

Multivariate Time Series Forecasting needs Cross Variable Loss

Sat, 01 Aug 2026 00:00:00 +0000

One paper accepted by SIGMOD 2027

Sat, 13 Jun 2026 00:00:00 +0000

“Algebraic Subgraph Counting” has been accepted by ACM SIGMOD 2027 (Research Round 1). Congratulations to Qiuyu.

OpenHalDet: A Unified Benchmark for Hallucination Detection across Diverse Generation Scenarios

Mon, 01 Jun 2026 00:00:00 +0000

OpenHalDet is a unified benchmark for hallucination detection across diverse generation scenarios of large language models. It brings together 17 datasets, 16 detection methods, and 5 backbone LLMs ranging from 3B to 70B parameters, contributed by researchers across Australia, the United States, the United Kingdom, and Singapore. The benchmark is released under the MIT license.

HYDRA: Proactive Android Malware Drift Adaptation via Hierarchical Graph Contrastive Learning

Wed, 20 May 2026 00:00:00 +0000

Concept drift, driven by the rapid evolution of Android malware, severely degrades the performance of machine learning detectors. Current adaptation strategies are often reactive, responding only after performance has dropped and imposing a significant manual annotation burden, or they are proactive but rely on unstable adversarial training and incomplete, single-level graph representations. To overcome these limitations, we propose HYDRA (Hybrid Drift Adaptation), a proactive adaptation framework that learns drift-invariant representations from hierarchically structured data. HYDRA first models applications using a hybrid graph structure, combining fine-grained Control Flow Graphs (CFGs) and coarse-grained Function Call Graphs (FCGs) to capture comprehensive behavioral patterns. It then introduces a novel cross-domain contrastive learning objective that aligns historical (source) and new (target) data distributions. By generating pseudo-labels for unlabeled target samples, our method pulls representations of semantically similar applications together, regardless of their domain, within a single, stable optimization process. This approach unifies feature learning and domain alignment, eliminating the need for complex adversarial objectives. Extensive experiments on large-scale, time-ordered malware datasets demonstrate that HYDRA achieves state-of-the-art performance with an F1-score of up to 96.9%, while reducing labeling effort by up to 87.5% compared to the best-performing baseline. Our work thus offers a robust and efficient solution to combat concept drift in security applications.

One paper accepted by ACM CCS 2026

Wed, 20 May 2026 00:00:00 +0000

“HYDRA: Proactive Android Malware Drift Adaptation via Hierarchical Graph Contrastive Learning” has been accepted by ACM CCS 2026. Congratulations to Han.

Agentic AI survey and ICDE 2026 tutorial

Fri, 15 May 2026 00:00:00 +0000

We released a comprehensive survey, “Data in Agentic AI”, and a corresponding ICDE 2026 tutorial, “Data-Centric Foundations of Agentic AI”. Congratulations to Yuxin.

Data in Agentic AI

Thu, 07 May 2026 00:00:00 +0000

A survey and living resource list on the data side of agentic AI: how these systems process data, retrieve it, remember it, coordinate over it, and fail because of it. 202 papers indexed so far, with an accompanying ICDE 2026 tutorial.

Data-Centric Foundations of Agentic AI

Fri, 01 May 2026 00:00:00 +0000

EXG: Self-Evolving Agents with Experience Graphs

Fri, 01 May 2026 00:00:00 +0000

GCCM: Enhancing Generative Graph Prediction via Contrastive Consistency Model

Fri, 01 May 2026 00:00:00 +0000

Multi-Agent Coordination Adaptation via Structure-Guided Orchestration

Fri, 01 May 2026 00:00:00 +0000

Towards Generative Graph Matching for Graph Edit Distance Computation

Fri, 01 May 2026 00:00:00 +0000

MGDN: A Graph of Graphs Neural Network for Malware Detection

Thu, 30 Apr 2026 00:00:00 +0000

Accelerating K-Core Computation in Temporal Graphs

Sun, 01 Mar 2026 00:00:00 +0000

Small Shifts, Large Gains: Unlocking Traditional TSP Heuristic Guided-Sampling via Unsupervised Neural Instance Modification

Sun, 01 Feb 2026 00:00:00 +0000

Algorithmic and Learning-based Approaches to Subgraph Matching and Counting: A Survey

Thu, 01 Jan 2026 00:00:00 +0000

Finding critical users in social networks with reinforcement learning

Thu, 01 Jan 2026 00:00:00 +0000

Mining Discriminative Salient Objects with Optimal Transport for Few-Shot Image Classification

Thu, 01 Jan 2026 00:00:00 +0000

RLMiner: Finding the Most Frequent k-sized Subgraph via Reinforcement Learning

Thu, 01 Jan 2026 00:00:00 +0000

WOCD: A semi-supervised method for overlapping community detection using weak cliques

Thu, 01 Jan 2026 00:00:00 +0000

Data in Agentic AI: A Comprehensive Survey

Wed, 31 Dec 2025 00:00:00 +0000

Give Me Some SALT: Structure-Aware Link Modeling for Temporal Weighted Link Prediction

Wed, 01 Oct 2025 00:00:00 +0000

Large language models meet text-attributed graphs: A survey of integration frameworks and applications

Wed, 01 Oct 2025 00:00:00 +0000

Towards Unsupervised Training of Matching-based Graph Edit Distance Solver via Preference-aware GAN

Fri, 19 Sep 2025 00:00:00 +0000

Machine Learning for Graph Data Management and Query Processing

Thu, 04 Sep 2025 00:00:00 +0000

HiGraph: A Large-Scale Hierarchical Graph Dataset for Malware Analysis

Tue, 02 Sep 2025 00:00:00 +0000

HiGraph is a large-scale hierarchical graph dataset for Android malware analysis. It contains 499,981 Android applications (50,661 malicious and 449,320 benign), where each application is represented by a two-level hierarchical structure combining Function Call Graphs (FCGs) and over 200 million embedded Control Flow Graphs (CFGs). The dataset spans 683 malware families collected between January 2012 and December 2022, enabling research on malware detection and classification, robustness to code obfuscation, and temporal evolution analysis of malware. Data and code are released under CC BY-NC-SA 4.0.

Efficient and Accurate Subgraph Counting: A Bottom-up Flow-learning based Approach

Fri, 01 Aug 2025 00:00:00 +0000

New Website Launched! 😃

Sun, 20 Jul 2025 00:00:00 +0000

We are thrilled to announce the official launch of the Data Intelligence Lab website! Explore our research, news, and publications.

Structure and Position-Aware Graph Modeling for Trajectory Similarity Computation Over Road Networks

Thu, 01 May 2025 00:00:00 +0000

Predicting Membrane Fouling of Submerged Membrane Bioreactor Wastewater Treatment Plants Using Machine Learning

Wed, 05 Mar 2025 00:00:00 +0000

AIGC for Graphs: Current Techniques and Future Trends

Tue, 04 Mar 2025 00:00:00 +0000

Covering K-Cliques in Billion-Scale Graphs

Sat, 01 Mar 2025 00:00:00 +0000

AI-Empowered Catalyst Discovery: A Survey from Classical Machine Learning Approaches to Large Language Models

Sat, 01 Feb 2025 00:00:00 +0000

Computing Historical k-Core in Parallel

Wed, 01 Jan 2025 00:00:00 +0000

IGFM: An Enhanced Graph Similarity Computation Method with Fine-Grained Analysis

Wed, 01 Jan 2025 00:00:00 +0000

Inferring gene regulatory networks by hypergraph generative model

Wed, 01 Jan 2025 00:00:00 +0000

Summary We present hypergraph variational autoencoder (HyperG-VAE), a Bayesian deep generative model that leverages hypergraph representation to model single-cell RNA sequencing (scRNA-seq) data. The model features a cell encoder with a structural equation model to account for cellular heterogeneity and construct gene regulatory networks (GRNs) alongside a gene encoder using hypergraph self-attention to identify gene modules. The synergistic optimization of encoders via a decoder improves GRN inference, single-cell clustering, and data visualization, as validated by benchmarks. HyperG-VAE effectively uncovers gene regulation patterns and demonstrates robustness in downstream analyses, as shown in B cell development data from bone marrow. Gene set enrichment analysis of overlapping genes in predicted GRNs confirms the gene encoder’s role in refining GRN inference. Offering an efficient solution for scRNA-seq analysis and GRN construction, HyperG-VAE also holds the potential for extending GRN modeling to temporal and multimodal single-cell omics.

RIDA: a robust attack framework on incomplete graphs

Wed, 01 Jan 2025 00:00:00 +0000

Bridging Large Language Models and Graph Structure Learning Models for Robust Representation Learning

Tue, 01 Oct 2024 00:00:00 +0000

Temporal Insights for Group-Based Fraud Detection on e-Commerce Platforms

Tue, 01 Oct 2024 00:00:00 +0000

Deep Learning Approaches for Similarity Computation: A Survey

Wed, 03 Jul 2024 00:00:00 +0000

Bipartite Graph Analytics: Current Techniques and Future Trends

Mon, 13 May 2024 00:00:00 +0000

Simple and Deep Graph Attention Networks

Tue, 19 Mar 2024 00:00:00 +0000

STG-Mamba: Spatial-Temporal Graph Learning via Selective State Space Model

Mon, 18 Mar 2024 00:00:00 +0000

Influence Maximization on Hypergraphs via Multi-Hop Influence Estimation

Tue, 16 Jan 2024 00:00:00 +0000

GQ: Towards Generalizable Deep Q-Learning for Steiner Tree in Graphs

Mon, 01 Jan 2024 00:00:00 +0000

TIformer: A Transformer-Based Framework for Time-Series Forecasting with Missing Data

Mon, 01 Jan 2024 00:00:00 +0000

FPGN: follower prediction framework for infectious disease prevention

Sat, 16 Sep 2023 00:00:00 +0000

Group-based Fraud Detection Network on e-Commerce Platforms

Fri, 04 Aug 2023 00:00:00 +0000

Denoising Variational Graph of Graphs Auto-Encoder for Predicting Structured Entity Interactions

Mon, 24 Jul 2023 00:00:00 +0000

Neural Similarity Search on Supergraph Containment

Thu, 25 May 2023 00:00:00 +0000

TMN: Trajectory Matching Networks for Learning Similarity Computation

Fri, 25 Mar 2022 00:00:00 +0000

Neural Subgraph Counting with Wasserstein Estimator

Tue, 08 Mar 2022 00:00:00 +0000

Reinforcement Learning Based Query Vertex Ordering Model for Subgraph Matching

Tue, 18 Jan 2022 00:00:00 +0000

Polarity-based Graph Neural Network for Sign Prediction in Signed Bipartite Graphs

Wed, 12 Jan 2022 00:00:00 +0000

Bipartite Graph Capsule Network

Tue, 11 Jan 2022 00:00:00 +0000

Powerful Graph of Graphs Neural Network for Structured Entity Analysis

Fri, 04 Jun 2021 00:00:00 +0000

Binarized Graph Neural Network

Thu, 08 Apr 2021 00:00:00 +0000

T3S: Effective Representation Learning for Trajectory Similarity Computation

Mon, 18 Jan 2021 00:00:00 +0000

EI-LSH: An early-termination driven I/O efficient incremental c-approximate nearest neighbor search

Wed, 30 Sep 2020 00:00:00 +0000

GoGNN: Graph of Graphs Neural Network for Predicting Structured Entity Interactions

Mon, 06 Apr 2020 00:00:00 +0000

I/O Efficient Algorithm for c-Approximate Furthest Neighbor Search in High-Dimensional Space

Wed, 01 Apr 2020 00:00:00 +0000

Skyline Nearest Neighbor Search on Multi-layer Graphs

Sat, 20 Apr 2019 00:00:00 +0000

I-LSH: I/O Efficient c-Approximate Nearest Neighbor Search in High-Dimensional Space

Sat, 06 Apr 2019 00:00:00 +0000